madwifi-ng (probably all versions) didnt compile on Gentoo and 2.6.24 kernel. It die with complain about ARCH x86 that doesnt match detected i386. This is due to the new 2.6.24 kernel and the merge of x86_64 and i386. Madwifi chech for ARCH in the scripts/get_arch.mk http://madwifi.org/browser/madwifi/trunk/scripts/get_arch.mk They expect i386 in that test : ARCH-$(CONFIG_X86) = i386 But gentoo portage, in toolchain-funcs.eclass set ARCH as x86 if the kernel 2.6.24 is present. Reproducible: Always
Do you have a patch? If not, then why are you filing bugs about issues affecting testing pre-releases of kernels to Gentoo bugzilla?
(In reply to comment #1) > Do you have a patch? Not really. As workarround I've changed the toolchain-funcs.eclass to export i386 as ARCH, but probably not a good idea. > If not, then why are you filing bugs about issues > affecting testing pre-releases of kernels to Gentoo bugzilla? Oh, was I wrong ? I suspect that will be an issue when 2.6.24 will goes out, and this will probably impact some other ebuild as well. Sorry if I was wrong to do it.
May or may not be wrong - but in either case, your "fix" is *definitely* wrong. This is something that has come up before. Ill take a look when I have caught up other bugs (unreleased == lower priority)
Created attachment 141779 [details, diff] madwifi-ng linux 2.6.24 i386/x86_64/x86 build fix
Comment on attachment 141779 [details, diff] madwifi-ng linux 2.6.24 i386/x86_64/x86 build fix I've made this patch for my x86_64 gentoo system. It should also work for i386.
Confirming this patch works on x86.
Well, the patch works, but I've never seen so many system freezes on my gentoo box... After some sniffing the driver dies and don't get up again, any iwconfig or any other wireless related commando just freeze, and if I restart or turn off the system, it freezes on the process and I have to do a hard reset.
(In reply to comment #8) > Well, the patch works, but I've never seen so many system freezes on my gentoo > box... > > After some sniffing the driver dies and don't get up again, any iwconfig or any > other wireless related commando just freeze, and if I restart or turn off the > system, it freezes on the process and I have to do a hard reset. > i confirm that, i have lot kernel data output when i try to connect my atheros card to cardbus socket :S lot kernel panicks :S
(In reply to comment #8) > Well, the patch works, but I've never seen so many system freezes on my gentoo > box... > > After some sniffing the driver dies and don't get up again, any iwconfig or any > other wireless related commando just freeze, and if I restart or turn off the > system, it freezes on the process and I have to do a hard reset. > update try to compile http://snapshots.madwifi.org/madwifi-ng-current.tar.gz it seems fixed in svn svn driver works for me fine:)
Well apart from lots of "Unknown sysctl binary path" error messages in dmesg, 0.9.3.3 works fine for me. There are no error messages in dmesg when using svn version.
*** Bug 207952 has been marked as a duplicate of this bug. ***
I am not going to use this patch in the tree, although thank you for your work on it - Upstream is waiting on 2 more patches before they release 0.9.4 - if you want you can test it without those 2 specific patches currently, one is for power saving mode, and the other is for retries when injecting packets. If you don't need either of those features, feel free to checkout the 0.9.4 branch from subversion, unfortunately, there is no timeframe on the two patches, the injection one was acked, but the powersave one, the author would like people to take a closer look at.
well, i've got some strange behaviour with this package: with thickless enabled it builds and works fine but when removing thickless from kernel it won't compile. i think that there's something wrong with the powermanagement that thickless fixes.
*** Bug 208513 has been marked as a duplicate of this bug. ***
Created attachment 142491 [details] new ebuild
Created attachment 142493 [details, diff] my getarch patch for 2.6.24 kernel
(In reply to comment #17) > Created an attachment (id=142493) [edit] > my getarch patch for 2.6.24 kernel > I tried Oleg's patch and ebuild and it did build. Now when I try to unload the ath_pci module, it segfaults. When I try to reboot, it gets stuck at bringing down the other network interfaces (eth0 and lo). Just wanted to see if anybody else experiences this.
(In reply to comment #18) > I tried Oleg's patch and ebuild and it did build. Now when I try to unload the > ath_pci module, it segfaults. When I try to reboot, it gets stuck at bringing > down the other network interfaces (eth0 and lo). > > Just wanted to see if anybody else experiences this. madwifi just releases the 0.9.4 [1] that is the one that work with 2.6.24. The 0.9.3.3 was not supported on .24 series. I'm using the svn version and it work like a charm on my ThinkPad. [1] http://madwifi.org/wiki/news/20080213/release-0-9-4-available
0.9.4 was released around 12:30 last night, I haven't gotten a chance to look into it/get it into the tree yet. No need to apply these patches and so forth, just give me a little bit to get it in there, unfortunately, 0.9.4 is based on 0.9.3.3 and not on trunk, so there is no macbook support, nor AR5007 (Eeepc) so, yeah...
(In reply to comment #19) > (In reply to comment #18) > > I tried Oleg's patch and ebuild and it did build. Now when I try to unload the > > ath_pci module, it segfaults. When I try to reboot, it gets stuck at bringing > > down the other network interfaces (eth0 and lo). > > > > Just wanted to see if anybody else experiences this. > > madwifi just releases the 0.9.4 [1] that is the one that work with 2.6.24. The > 0.9.3.3 was not supported on .24 series. > > I'm using the svn version and it work like a charm on my ThinkPad. > > [1] > http://madwifi.org/wiki/news/20080213/release-0-9-4-available > Alex: Are you using an ebuild from bugzilla?
(In reply to comment #20) > 0.9.4 was released around 12:30 last night, I haven't gotten a chance to look > into it/get it into the tree yet. No need to apply these patches and so forth, > just give me a little bit to get it in there, unfortunately, 0.9.4 is based on > 0.9.3.3 and not on trunk, so there is no macbook support, nor AR5007 (Eeepc) > so, yeah... Sorry Steve. I didn't know 0.9.4 was only released last night. I just tried the 2.6.24 kernel last night and this issue came up. Thanks for the info.
(In reply to comment #21) > > Alex: Are you using an ebuild from bugzilla? > Yes, see bug #193549
Should 0.9.4 work with AR5006EG?
(In reply to comment #24) > Should 0.9.4 work with AR5006EG? > Take a look on the compatibility list [1], but as I can read, it should work in 0.9.3.3 also. In release notes [2], you will see Known issues that list AR5007 and AR5008 as not supported. [1] http://madwifi.org/wiki/Compatibility/Atheros#AtherosAR5006EG [2] http://madwifi.org/wiki/Releases/0.9.4
This ticket outlines the problem: http://madwifi.org/ticket/859 but I see that it was updated about a week ago with this: "lspci may mis-identify 168c:001c AR5007 chipsets as AR5006EG." Which sounds like good news to me.
I've commited madwifi-ng-tools-0.9.4 as well as madwifi-ng-0.9.4 just now. Please let me know how it works for you.
I still get the following in dmesg so madwifi still isn't working here: wifi%d: unable to attach hardware: 'Hardware revision not supported' (HAL status 13) The ebuilds built just fine though. Thank you very much.
reopening as this is not fixed in the stable tree
Except that 2.6.24 isn't stable...
Thanks for the commit, Steve. Everything's working fine here with tuxonice 2.6.24-r2 kernel and AR5006EG chipset.
Steev, thanks for fast commit. Still using 2.6.20-2.6.23 series kernels here but seems to work fine as well. Glad to see upstream finally include minstrel rate module.
quick reminder here; 2.6.24 is going stable very soon
Well folks, 2.6.24-gentoo-r3 went stable now and I ran into this issue while updating my stable system. Maybe you should consider to stabilize madwifi-ng-{,tools-}0.9.4 as well. Thanks. Dustin
I have this problem on amd64. Stable madwifi-ng does not compile with stable gentoo-sources. Not good.
I just emerged successfully madwifi-ng-0.9.4 and madwifi-ng-tools-0.9.4. It seems to run just fine with a 2.6.24-gentoo-r3 kernel on AMD64. Here's the syslog info: ath_hal: module license 'Proprietary' taints kernel. ath_hal: 0.9.18.0 (AR5210, AR5211, AR5212, RF5111, RF5112, RF2413, RF5413) wlan: 0.9.4 ath_pci: no version for "ieee80211_encap" found: kernel tainted. ath_pci: 0.9.4 PCI: Enabling device 0000:02:00.0 (0000 -> 0002) ACPI: PCI Interrupt 0000:02:00.0[A] -> GSI 17 (level, low) -> IRQ 17 ath_rate_sample: 1.2 (0.9.4) wifi0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps wifi0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps wifi0: turboG rates: 6Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps wifi0: H/W encryption support: WEP AES AES_CCM TKIP wifi0: mac 5.6 phy 4.1 radio 4.6 wifi0: Use hw queue 1 for WME_AC_BE traffic wifi0: Use hw queue 0 for WME_AC_BK traffic wifi0: Use hw queue 2 for WME_AC_VI traffic wifi0: Use hw queue 3 for WME_AC_VO traffic wifi0: Use hw queue 8 for CAB traffic wifi0: Use hw queue 9 for beacons wifi0: Atheros 5212: mem=0x84000000, irq=17
madwifi-ng-0.9.4 and madwifi-ng-tools-0.9.4 compiles + runs fine with 2.6.24-gentoo-r3 at an IBM ThinkPad T41 (x86) : tfoerste@n22 ~ $ uname -a Linux n22 2.6.24-gentoo-r3 #5 Tue Mar 18 19:26:55 CET 2008 i686 Intel(R) Pentium(R) M processor 1700MHz GenuineIntel GNU/Linux
I report that here, running 2.6.24, madwifi-ng-0.9.4 hangs the system when using master mode (AP). When I get more information I'll comment over here. Thanks Linux smash64 2.6.24-gentoo-r4 #4 SMP Thu Apr 24 19:38:16 CEST 2008 x86_64 AMD Athlon(tm) 64 X2 Dual Core Processor 3800+ AuthenticAMD GNU/Linux
On a related note, it may be worth pointing out that the latest stable x86 ebuild of madwifi-ng (0.9.3.3) doesn't build with latest stable x86 ebuild of gentoo-sources (2.6.24-r8). It would be good to have madwifi-ng(-tools)-0.9.4, or newer if necessary, marked as stable as soon as possible on x86 and other affected architectures.
(In reply to comment #39) > On a related note, it may be worth pointing out that the latest stable x86 > ebuild of madwifi-ng (0.9.3.3) doesn't build with latest stable x86 ebuild of > gentoo-sources (2.6.24-r8). It would be good to have madwifi-ng(-tools)-0.9.4, > or newer if necessary, marked as stable as soon as possible on x86 and other > affected architectures. > I confirm on my recent 2.6.24-r8.
Since ath5k is in kernel from some times and Madwifi team saying it will not develop new features, could users switch to the in-kernel ath5k driver in a near futur ? From the MadWifi website [1] : "MadWifi has been declared legacy, in response to our commitment to ath5k. In the long run ath5k will replace MadWifi. For the time being MadWifi will still be supported, bugs will get fixed and HAL updates will be applied where possible. But it becomes unlikely that we'll see new features or go through major changes on that codebase. The driver is mature and stable, though, and works well for most users. For now you should choose MadWifi if you depend on solid WLAN support. But we also encourage all interested users to give ath5k a try." [1]: http://madwifi-project.org/wiki/About/MadWifi
Hi Alexandre, While it would be nice to have everyone switch to ath5k in kernel, there are still some issues (minor) with it, sometimes it can be hit or miss (same can be said about madwifi-ng though ;) ) with support and networking in general. I will continue to push out new releases when the Madwifi team releases them, and since I use the injection feature of the madwifi-ng drivers the in-kernel will not work for me (yet.) I've tried the patches to the mac80211 stack itself for monitor mode support but have only had limited success with them. I noticed a 2.6.28 release of the kernel, I initially thought perhaps madwifi-ng wasn't building against it when you reopened this ticket. I am glad that isn't the case. Perhaps the next time there is a madwifi release, I will add an einfo suggesting users try out the ath5k in kernel.
The last time I tried ath5k it wouldn't work in master mode. Has that been fixed? Neither ath5k or madwifi-ng work with my laptop's AR5006EG.
http://madwifi-project.org/wiki/Compatibility/Atheros#AtherosAR5006EG(In reply to comment #43) > The last time I tried ath5k it wouldn't work in master mode. Has that been > fixed? Neither ath5k or madwifi-ng work with my laptop's AR5006EG. > This chip seems to be supported by madwifi : http://madwifi-project.org/wiki/Compatibility/Atheros#AtherosAR5006EG Be sure to have no conflicts with ath5k (kernel) and ath5k_pci (madwifi). https://bugs.launchpad.net/ubuntu/+bug/286861
(In reply to comment #42) > Hi Alexandre, > > While it would be nice to have everyone switch to ath5k in kernel, there are > still some issues (minor) with it, sometimes it can be hit or miss (same can be > said about madwifi-ng though ;) ) with support and networking in general. I > will continue to push out new releases when the Madwifi team releases them, and > since I use the injection feature of the madwifi-ng drivers the in-kernel will > not work for me (yet.) I've tried the patches to the mac80211 stack itself for > monitor mode support but have only had limited success with them. I noticed a > 2.6.28 release of the kernel, I initially thought perhaps madwifi-ng wasn't > building against it when you reopened this ticket. I am glad that isn't the > case. Perhaps the next time there is a madwifi release, I will add an einfo > suggesting users try out the ath5k in kernel. > I've not reopened this bug ;) This was done by dsd@g.o on 2008-02-14. As I can understand, we still need both ath5k and madwifi, but expect conflicts between in-kernel driver and madwifi one, since they share same hardware. Could this be added as a warning in madwifi ebuild ?
